Increasing evidence suggests that the environment is related to many public health challenges. Unequal distributions of services and resources needed for healthy lifestyles may contribute to increasing levels of health disparity. However, empirical studies are not sufficient to understand the relationship between health disparity and the built environment. This dissertation examines how health disparity are associated with the built environment and if the environmental conditions that support physical activity and healthy diet are associated with lower health disparity. This research uses a multidisciplinary approach, drawing from urban planning, regional economics and public health. The data came from the Behavioral Risk Factor Surveillance System, and the GIS derived environmental data and the 608-respondent survey data from a larger study conducted in urbanized King County, Washington. Health disparity was measured with the Gini-coefficient, and health status and obesity were used as indicators of health. Hot spot analysis was used to identify the spatial aggregations of high health disparity, and multiple regression models identified the environmental correlates of health disparity. The overall trend showed that disparity has increased in most states in the US over the past decade and the southern states showed the highest disparity levels. Strong spatial autocorrelations were found for disparities, indicating that disparity levels are not equally distributed across different geographic areas. From the multivariate analyses estimating disparity levels, spatial regression models significantly improved the overall model fit compared to the ordinary least-square models. Areas with more supportive built environments for physical activity had lower health disparities, including proximity to downtown (+) and access to parks (+), day care centers (+), offices (+), schools (+), theaters (+), big box shopping centers (-), and libraries (-). Overall results showed that the built environment, compared to the personal factors, was more strongly correlated with health disparities. This study brings attention to the problem of health disparity in the US, and provides evidence supporting the existence of spatial disparity in the environmental support for a healthy lifestyle. Further research is needed to better understand environmental and socioeconomic conditions associated with health disparity among more diverse population groups and in different environmental settings.

No description available.

In the present paper, we examine the effect of socioeconomic characteristics on the life expectancy of men and women in the Stockholm metropolitan area. Detailed individual data allows for a novel approach where observations can be displayed in high resolution. As is often the case with geographical data, the variables display high spatial autocorrelations, which imply that observations in proximity are more, or less, similar than what could be expected under the assumption of independent and identically distributed observations. Presence of spatial autocorrelation makes conventional regression models nonfunctional, and a model that accounts for this is therefore specified. In addition, a distance-band which reflects the distance and association between observations is determined. Lagrange Multiplier tests, AIC, log-likelihood, and the Schwarz criterion suggest that a spatial error model with a 300-meter distance band is appropriate for the data at hand. The findings suggest that: (1) Belonging to a minority group has the strongest effect on life expectancies and (2) the effect is negative for both genders, although the negative impact is stronger for males. Tests for spatial autocorrelation on the residuals suggest that the adopted spatial error model captures nearly all spatial autocorrelation in the data, compared to alternative models.

No description available.

京都大学 / 新制・課程博士 / 博士(地球環境学) / 甲第24954号 / 地環博第245号 / 新制||地環||49(附属図書館) / 京都大学大学院地球環境学舎環境マネジメント専攻 / (主査)教授 星野 敏, 准教授 鬼塚 健一郎, 教授 伊藤 孝行 / 学位規則第4条第1項該当 / Doctor of Global Environmental Studies / Kyoto University / DFAM

No description available.

Financial inclusion has gained recognition as both a domestic and international governance objective. However, full participation in the financial sector remains an elusive goal, and a number of significant questions present themselves regarding defining the scope of financial inclusion and formulating efficacious policy to ensure access to and promoting the usage of financial services. Paramount among these questions is the relationship between the geographic aspects of retail financial markets and consumer outcomes including rates of savings and indebtedness, the types of consumer credit utilized, and levels of unbanked and underbanked populations. The central aim of this research is to address this lack of understanding by using quantitative analytical tools including geographic information systems (GIS) and spatial regression analysis to examine relationships between the uneven geography of retail financial services, mortgage lending activity, and sociodemographic variables. Four metropolitan study areas in the United States—Las Vegas, Nevada; Los Angeles, California; Miami, Florida; and Philadelphia, Pennsylvania—are examined in order to address a range of question related to the neighborhood level determinants of financial inclusion. This study will provide a foundation for improving policy solutions through contributing to the understanding of how data-driven and analytical approaches can be applied to this problem. Specifically, the following research questions are addressed: 1) How does the spatial distribution of mainstream financial institutions (banks and credit unions) and alternative financial service providers (AFSPs) contribute to financial inclusion at the neighborhood level? What is the geographic relationship between these services; and how does access to these services interact with neighborhood demographic variables and mortgage lending activity? 2) How can traditional approaches to spatial analysis of mortgage lending be improved and expanded to incorporate new spatial analysis methods and better understand how mortgage credit denial and subprime lending interact with one another, as well as with neighborhood demographic variables? Building on scholarship in the academic areas of community reinvestment, asset building, and economic geography, this research contributes a number of new insights and refinements in methodology. The results of spatial regression analyses reveal significant predictive relationships, even after controlling for sociodemographic variables and spatial clustering by using simultaneous autoregressive (SAR) models. This research is unique in its examination of the relationship between the landscape of financial services in neighborhoods and mortgage lending activity, and finds that increasing levels of subprime mortgage lending in neighborhoods is predictive of nearer distance to AFSPs. Another finding is that higher percentages of black and Latino populations in neighborhoods are predictive of nearer proximity to AFSPs and greater distances to mainstream brick-and-mortar financial institution locations. A new method is developed to address the spatial void hypothesis, the spatial relationship between mainstream financial institutions and AFSPs. The results of binary logistic regression models indicate that neighborhoods where alternative service providers are more prevalent comparatively feature lower average income levels, higher percentages of minority residents, lower levels of educational attainment, and higher levels of both mortgage application denial and subprime mortgage lending. Advances are also made in developing regression models to address relationships between sociodemographic variables and mortgage lending activity. Using SAR modeling, this study finds that mortgage purchase denial is a strong predictor of subprime lending for home purchase and refinance loans. Confirming prior research findings with a new method, the percentage of the population that is black and Latino is found to be a statistically significant predictor of mortgage purchase denial, as well as rates of subprime mortgage purchase lending. / Geography

nÃo hà / This paper aims to estimate a residential water demand function for the city of Fortaleza (CearÃ), considering the potential impact of the spatial effects on water consumption. The analysis is developed from the investigation of presence of spatial autocorrelation in residential water consumption. For this, the tools of exploratory spatial data analysis (ESDA) were utilized. Subsequently, specific tests are performed to determine the sources of spatial autocorrelation, i.e., if the autocorrelation is caused by the spatial distribution of water consumption or by effects not modeled. Identified the sources of spatial autocorrelation, four water demand functions were estimated, which had as explanatory variables the average price, the difference, income, number of residents and the number of rooms, under different specifications. At first, we estimated a model without special effects; in the second, we estimated the specification of the spatial error model (SEM), which incorporates the spatial autocorrelation in the form of autocorrelation in the error terms; in the third, we estimated the spatial autoregressive model (SAR), where the spatial autocorrelation is incorporated through the spatial lag of the dependent variable; and finally, we estimated the spatial model autoregressive moving average (SARMA), which is the union of the two previous models. The results show that spatial autocorrelation exists in two forms (error and lag), indicating that the SARMA model is the most indicated to model the residential water demand in the city of Fortaleza, in contrast to suggested by Chang et al.(2010), House-Peters et al. (2010), Franczyk e Chang (2008), Ramachandran e Johnston (2011), which used the SEM model. It is concluded that it is important to consider the possibility of spatial effects in the estimation of a residential water demand function, once that not incorporate spatial effects in the analysis underestimate the effect of the variables average price and number of residents on residential water demand, while overestimating the effect of the variables income and number of rooms. / Esta dissertaÃÃo tem como objetivo estimar uma funÃÃo de demanda residencial por Ãgua para a cidade de Fortaleza (CearÃ), considerando o provÃvel impacto do efeito espacial no consumo de Ãgua. A anÃlise se desenvolve a partir da investigaÃÃo a respeito da presenÃa de autocorrelaÃÃo espacial no consumo residencial de Ãgua. Para tal, foram utilizadas as tÃcnicas de anÃlise exploratÃria espacial de dados (ESDA). Posteriormente, sÃo realizados testes especÃficos para determinar as fontes da autocorrelaÃÃo espacial, ou seja, identificar se a autocorrelaÃÃo à causada pela distribuiÃÃo espacial do consumo de Ãgua ou pelos efeitos nÃo modelados. Identificadas as fontes de autocorrelaÃÃo espacial, foram estimadas quatro funÃÃes de demanda de Ãgua, que tinham como variÃveis explicativas o preÃo mÃdio, a diferenÃa, a renda, o nÃmero de residentes e o nÃmero de cÃmodos, sob diferentes especificaÃÃes. Na primeira, utilizou-se um modelo sem efeitos espaciais; na segunda, utilizou-se a especificaÃÃo do modelo de erros espaciais (SEM), que incorpora a autocorrelaÃÃo espacial na forma de autocorrelaÃÃo nos termos de erro; na terceira, utilizou-se o modelo espacial autorregressivo (SAR), onde a autocorrelaÃÃo espacial à incorporada atravÃs da defasagem espacial da variÃvel dependente; e por Ãltimo, utilizou-se o modelo espacial autorregressivo de mÃdias mÃveis (SARMA), que à a uniÃo dos dois modelos anteriores. Os resultados mostram que existe autocorrelaÃÃo espacial nas duas formas (erro e defasagem), indicando que o modelo SARMA à o mais adequado para modelar a demanda residencial por Ãgua na cidade de Fortaleza, ao contrÃrio do proposto por Chang et al. (2010), House-Peters et al. (2010), Franczyk e Chang (2008), Ramachandran e Johnston (2011), que utilizaram o modelo SEM. Conclui-se, portanto, que à importante levar em consideraÃÃo a possibilidade de efeitos espaciais na estimaÃÃo de uma funÃÃo de demanda residencial por Ãgua, na medida que a nÃo incorporaÃÃo dos efeitos espaciais subestima o efeito das variÃveis preÃo mÃdio e nÃmero de residentes sobre a quantidade consumida de Ãgua, enquanto superestima o efeito das variÃveis renda e nÃmero de cÃmodos.

Hedonic pricing methods and GIS (Geographic Information Systems) were used to evaluate relationships between sale price of single family homes and landscape fragmentation and natural land cover. Spatial regression analyses found that sale prices increase as landscapes become less fragmented and the amount of natural land cover around a home increases. The projected growth in population and employment in the Richmond, Virginia region and subsequent increases in land development and landscape fragmentation presents a challenge to sustaining intact healthy ecosystems in the Richmond region. Spatial regression analyses helped illuminate how land cover patterns influence sale prices and landscape patterns that are economically and ecologically advantageous.

Identifying high-risk areas for child maltreatment to ultimately aid public health agencies for interventions is necessary for protecting children at high risk. Rates of substantiated neglect and physical/emotional abuse in 2000-2002 are computed for the census tracts in the urban area of five counties in Metro Atlanta, Georgia, and analyzed using spatial regression to determine their relationships with twelve risk variables computed from the Vital Records births and the 2000 Census data. After accounting for multicollinearity among risk variables and spatial autocorrelation among observations for neighboring locations, it is found that high percentages of (1) births to non-married mothers, (2) births to mothers who smoked or drank alcohol during pregnancy, (3) unemployed males and females, and (4) single-parent families with children under age six best predict the rates of substantiated neglect, and that high percentage of births to mothers who smoked or drank alcohol during pregnancy best predicts the rates of substantiated physical/emotional abuse.